K66 Beta Test

Status
Not open for further replies.
For anybody else playing Kerbal Space Program have confirmed that KSPIO works with Teensy 3.5 if you add packed attribute to the vessel data struct:
struct __attribute__((__packed__)) VesselData
As part of using this working on a graphical artificial horizon mapping a sphere. For those playing with beta units who have a display handy the 'renderer' for a wireframe ball looks like
Code:
int stepsize =30;
int scale = 24;
int center = 25;
float pitch = 30;
float roll =0;
float heading =0;

float zpoint(float longditude,float latitude) //returns depth information, used to mask rear of sphere
{
  return 1*cos(longditude+heading )*abs(cos(latitude))*cos(pitch)+sin(pitch)*-sin(latitude);
}
float xpoint(float longditude,float latitude)  //get x values from lat/long/heading
{  
  return scale*sin(longditude+heading)*cos(latitude);
}
float ypoint(float longditude,float latitude)  //get y value from lat/long/heading and pitch
{
  return scale*cos(longditude+heading )*sin(pitch)*cos(latitude)+scale*sin(latitude)*cos(pitch);  
}
//
void line(float x1, float y1,float x2,float y2) //line drawing - change to suit display library
{
  display.drawLine(floor(x1), floor(y1), floor(x2),floor(y2) , BLACK);
}
//
void placeLine(float startLong,float startLat,float endLong,float endLat)
{ //applies roll to final co-ordinates via vector conversion
  if (zpoint(startLong,startLat)>0&&zpoint(endLong,endLat)>0){
    float startDistanceFromCenter = sqrt(sq(xpoint(startLong,startLat))+sq(ypoint(startLong,startLat)));
    float endDistanceFromCenter = sqrt(sq(xpoint(endLong,endLat))+sq(ypoint(endLong,endLat)));
    float startAngle = atan2(ypoint(startLong,startLat),xpoint(startLong,startLat));
    float endAngle = atan2(ypoint(endLong,endLat),xpoint(endLong,endLat));  
    int rotatedXStart = center+floor(startDistanceFromCenter*cos(roll+startAngle ));
    int rotatedYStart = center+floor(startDistanceFromCenter*sin(roll+startAngle ));
    int rotatedXEnd = center+floor(endDistanceFromCenter*cos(roll+endAngle));
    int rotatedYEnd = center+floor(endDistanceFromCenter*sin(roll+endAngle));
    line(rotatedXStart, rotatedYStart, rotatedXEnd,rotatedYEnd );
  }  
}
void drawBall(float pitchIn,float headingIn,float rollIn) //produces geometry
{  
 pitch=pitchIn; 
 heading=headingIn; 
 roll = rollIn;
  for (int i = 0; i < floor(360/stepsize); i++) {
        float angle = float(i*stepsize)/180*PI;
        int meridiansize = floor(90/stepsize);
        for (int l = -meridiansize; l< meridiansize+1; l++) {    
        float currantLat = float(l)*stepsize/180*PI;  
        placeLine(angle,currantLat,angle+PI/180*stepsize,currantLat);
        placeLine(angle,currantLat,angle,currantLat+PI/180*stepsize);
      }     
  }
}

void loop()
{
    display.clearDisplay(); 
  drawBall(sin(float(millis())/4000),float(millis())/2000,sin(float(millis())/2000)/2);
    display.display();
}
Please note this is BYO display library so add your own setup code and mod the 'clear' and 'display' elements of loop and the 'line' procedure to suit.

Lots of optimization options since do not think this is actually deploying the floating point hardware, and I'm sure there is a way to apply roll without the vector conversion stuffed in the placeLine there but my trig isn't strong enough to make things happen in one go.
 
This is an extremely minor nit. I'm sure there was a good reason for it, but I was just surprised a little at the 2nd USB header is offset a little.

I was soldering in some headers on my beta test Teensy 3.6 (40 pin female stacking headers cut down to 24 on the outside, 3 female 5-pin headers for the inside pins, with one pin removed from one of the 5 pin headers). I don't really have good jigs and such for soldering headers to boards, so I tend to just use male headers to cross brace the female headers, and to keep the spacing correct. I solder in the outside headers, no problem. I solder in the Vusb, aref, A10, A11 header, no problem after removing the one pin. I solder in the program, reset, 3.3v, ground, Vbattery header no problem (after I use a rotary tool to shave down the header).

Then I go to solder in the last header (secondary USB header), and I notice I can't use a row of male header pins between the outer headers, and the other inner header to keep the 2nd USB in place when I'm soldering it. In the grand scheme of things it is minor, but I tend to see things in through hole soldering to always be lined up in 0.1" rows, so that you could attach things to protoboards. I just taped the 2nd header to the first to keep it in place when soldering it.

I also noted that there are solder pads near where the pins for the secondary USB header goes, and you have to be careful not to bridge the gap between the USB pins and the solder pads. Probably 6 months ago, I would not have had the control to do it the first time, without having to use solder wick to clean things up (I still tend to glob my solder connections, but they are now smaller globs).

I looked around and I did not see the two holes for soldering the 32k real time clock crystal, even though there is a pin for Vbattery. I imagine we will learn more tonight when the kickstarter campaign goes live, and pictures can be posted.

Looking at the underneath pads (6 sets of pads, 3 on each side), it may be an interesting challenge to bring all of these pins out via PCBs like the FrankB/TallDog boards do for 3.2. Particularly since the Teensy 3.6 is already an inch longer.
 
Yes it looked to me also like the USB host pins are offset. In particular I believe they are half way between two pins in one direction and about 1.5 pin spaces in the other direction...

Again can say more in hopefully under an hour :D

I am playing around with a quick and dirty board design, that probably won't be like tall dogs as I personally like having lots of three pin like connections with power and ground. Right now I am thinking of doing those for most of the pins on outside of board and maybe a few others. But then have room to maybe simply bring out the others near where they are on the board, but not with three pin connections...

Trying to decide how much else I want on first experiment board. Things like AX servo, or some other stuff, like quick and dirty sound and LEDS or simply add that through (wit or witout ) boards on top...

I do have a few other questions as well, but will wait until we can post pictures...
 
@MM "I looked around and I did not see the two holes for soldering the 32k real time clock crystal"
> The RTC code I tried works - it seems there is a Factory included RTC crystal - and no user mount holes. I was going to ask the other day about the tolerance chosen for the crystal.

I'll try to keep post_8 updated if anyone has good summary info for 'getting started' - PM me if you put up good pictures or samples or see a telling post. Or a nice color coded pinout card. I have the one I like - WORD table on 0.1" spacing to hold up to the T_3.6 'too many to count' pins - but it may not be perfectly up to date.
 
I have it up - not seeing it yet:
https://www.kickstarter.com/discover/advanced?term=teensy&category_id=16&sort=magic&seed=2452410&page=1

Not sure if Paul with post a T_3 UPDATE?
https://www.kickstarter.com/projects/paulstoffregen/teensy-30-32-bit-arm-cortex-m4-usable-in-arduino-a/description

@MM: 8 am ...
Quick update - we're planning to have the Kickstarter campaign go live tomorrow (Aug 17) at 8 am pacific time.

The moment the Kickstarter page is up, it'll be ok to post pictures, even share on social media. Just about 15 hours to go...
 
I can get to kickstarter.com at 11:05am EST (USA). No teensy yet, but I don't expect it for another 12 hours or so.
Actually from Paul's quote:
Quick update - we're planning to have the Kickstarter campaign go live tomorrow (Aug 17) at 8 am pacific time.

The moment the Kickstarter page is up, it'll be ok to post pictures, even share on social media. Just about 15 hours to go...

So should be soon!

It is alive: https://www.kickstarter.com/projects/paulstoffregen/teensy-35-and-36?ref=discovery
 
I wasn't first: You and 2 other people supported this project.

I backed - now I need to read and see what the KS says :)

edit: - 35 minutes in ... I hit refresh this time
20 backers
$1,789 pledged of $5,000 goal
21 days to go
 
Last edited:
Since it is live now, here is my Excel document. The first tab is like the card, and the two other tabs show the ALT functions for each of the pins for T3.5 and T3.6. Hopefully it is reasonably correct.

If you do find issues, please let me know.

Note: Had to upload as a zip file...
 

Attachments

  • Teensy K66 Beta Pinout.zip
    22.6 KB · Views: 364
And it is up to $4,191. I imagine it will reach $5,000 by the end of today, but even if the initial orders taper off, it should easily make the goal.
 
And ordered(!) Now that mere mortals can see the pictures it looks like the Teensy 3.5 and 3.6 share a common PCB but are fitted with different components. Distinguishing between the two is going to be problematic - can guarantee that I will get it wrong!. Maybe put a small sticker on the SD card connector?

Also the Kickstarter is missing pjrc.com swag! I thought that was mandatory for all Kickstarter projects. Where are the stickers, morale patches, t-shirts, coffee cups and hats? :)
 
Nice ... not days needed but minutes ... to achieve the goal! Congratulations, Paul, this speeks for itself! :cool:

... and counting!
 
Actually I wish I could combine a few of the kickstarter rewards. Example would like 2 of the 3.5 boards as to try out the 5v stuff, as well as multiple 3.6 boards... But hopefully can order more soon when they are listed on the PJRC website!
 
Hmmm, it doesn't mention touch sensing inputs in the Teensy 3.5.

And I have to imagine that in the November time frame, you will be able to order more directly from PJRC, and probably Adafruit, Sparkfun, etc.
 
It didn't quite make the $5K goal in an hour : 60 backers $4,465 pledged of $5,000 goal

But at about 75 minutes : 74 backers $5,187 pledged of $5,000 goal

Interesting to see Paul did T_3.5 in parallel - it seemed it was going to be identical and might show up like that.

Looking at KS:

Features specific to Teensy 3.5:
> does not show # Touch Sensing Inputs

Features specific to Teensy 3.6:
> 11 Touch Sensing Inputs

They seem to share a PCB - with internal pins for : USB High Speed (480 Mbit/sec) Port
> Assume those are not connected on the Teensy 3.5? Or will they have some alternate functional connectivity?


edit - noticed all KS rewards are limited quantity - I wonder which won't hit their limit?
 
Last edited:
Status
Not open for further replies.
Back
Top